Papers with Unsupervised automatic readability assessment
Zero-shot Large Language Models for Automatic Readability Assessment (2026.acl-long)
Copied to clipboard
| Challenge: | Existing unsupervised automatic readability assessment methods have important practical and research applications. |
| Approach: | They propose a new zero-shot prompting methodology for automatic readability assessment (ARA) they propose combining large language models with readability formula scores to improve robustness . |
| Outcome: | The proposed method outperforms prior methods on 13 of 14 datasets and improves on contextual and shallow features of readability. |